Designing and maintaining large-scale ETL pipelines on on-prem systems, integrating and transforming hundreds of terabytes of data from global enterprise clients using Microsoft SQL Server and Python.
Optimizing processing workflows for several high-volume clients, reducing multi-hour data processing jobs to under 1 hour.
Developing a VS Code extension adopted by 30+ engineers to provide SQL-query shortcuts and automation, improving developer productivity and reducing repetitive query-writing time.
Software Engineering Intern
Jul. 2023 -- Dec. 2024
Evil GeniusesLos Angeles, CA
Built machine learning models used by coaches and management to evaluate over 1,000 professional players and support roster selection decisions.
Created a computer vision system (YOLOv5 + PyTorch) to detect logo placement and viewership across broadcast footage, providing analytics used in sponsorship negotiations and contributing to a $75,000 increase in annual sponsorship revenue.
Designed automated ETL pipelines using AWS Lambda, S3, and EC2 to ingest and normalize 2 TB/month of analytics data from 4 APIs into PostgreSQL, eliminating 10+ hours/week of manual data processing for staff.
Founder
Jan. 2021 -- Dec. 2023
Macrohard LLCChicago, IL
Built a real-time predictive analytics model for competitive gaming broadcasts, achieving 4% higher accuracy than all publicly available baselines and deployed in live productions.
Led a 7-person team across engineering, design, and operations; grew product revenue to $15,000 through sales and client partnerships.
Building an end-to-end platform (React web + Expo mobile) that authenticates users, stores location data, and visualizes their travel history across time.
Architecting a MySQL schema and data-ingestion pipeline for 30,000+ GPS points per user; optimized geospatial queries, reducing page load from 30s → 2s.
Implementing social features including friend connections and selective location-sharing, with secure token-based authentication and permission controls.
Developed a scalable real-time game simulation platform with 11,000 monthly users and 100–500 simulations per day, resulting in a contract with Riot Games to create a custom enterprise version.
Implemented a performant Node.js + MongoDB architecture supporting concurrent game sessions with low-latency updates over WebSockets.